1. Field of the Invention
The present invention relates to a packet transmission method and a packet transmission system for transmitting a packet via a network.
2. Description of the Related Art
In recent years, demand for establishing a highly reliable network for transmitting packets via Ethernet or the like with high reliability has risen. In addition, demand for a network capable of continuing communication without packet loss when a communication failure occurs to a packet network has risen. Generally, as a method of preventing instantaneous interruption during occurrence of a failure, there is proposed a method of switching an active device to a backup device without instantaneous interruption. However, to execute such a method, it is necessary to incorporate various functions such as a function of always matching a phase of traffic of the active device to that of the backup device and a function of a mechanism for detecting a failure and a congestion of the active device in a system.
The invention related to the present invention is disclosed in Japanese Patent Application Laid-Open No. 01-231542.
To continue traffic transmission without interruption when a failure occurs, a method of carrying traffic from a transmitting side to two paths, i.e., a path to the active device and a path to the backup device, and of causing a receiving side to select one of traffics from the active and backup devices and to output the selected traffic is adopted. With this method, one of the active and backup devices that has a smaller delay is matched to the other device that has a greater delay. Namely, the traffic from the device having the smaller delay is kept waiting for arrival of the traffic from the device having the greater delay so as to enable the same traffic to be output from the active and backup devices at the same timing (match traffic phases to each other). Furthermore, the traffic from the active device is always selected and output. If a failure occurs to the active device, then the receiving side detects occurrence of the failure to the active device by some means, switch selection from the active device to the backup device, and causes the backup device to output the traffic. The conventional technique has, however, the following disadvantages.
To prevent occurrence of interruption of traffic when the active device is switched to the backup device, it is disadvantageously necessary to control matching packet arrival timing between the active device and the backup device. To perform the control, a buffer that absorbs a transmission delay of the active device and that of the backup device is necessary. Furthermore, the greater delay is disadvantageously and always adopted as the transmission delay.